The emergency department (ED) serves as our society's medical safety-net, especially for those patients with critical illnesses. An accurate and reliable clinical prediction instrument applied to a broad range of seriously-ill ED patients to estimate mortality would be extremely useful. An instrument of this nature may also be used to compare quality of care, improve resource allocation, and select patients for aggressive hemodynamic monitoring and therapy. Currently, there is a paucity of effective tools used for predicting mortality in seriously-ill ED patients. The Mortality in Emergency Department Sepsis (MEDS) Score attempted to fill this gap, but has several limitations and thus does not effectively translate into clinical practice. The first hypothesis of the investigations proposed here is that ED patient and laboratory data will be used to derive a novel mortality risk score among patients admitted to the hospital with systemic inflammatory response syndrome (SIRS). The second hypothesis is that the resultant mortality risk score will stratify ED patients into distinct risk strata corresponding to increasing levels of mortality. The third hypothesis is that the resultant mortality risk score will be superior to the MEDS Score when applied to ED patients admitted to the hospital with SIRS. The fourth hypothesis is that the mortality risk score and its individual components will be highly reliable when applied by physicians. The Specific Aims therefore include: (1) performing a multi-center prospective ED cohort study to derive a mortality risk score using multivariable logistic regression analysis;(2) performing a multi-center prospective ED cohort study to validate the mortality risk score derived in Specific Aim 1 and to compare its performance with that of the MEDS Score in the same ED population;and (3) performing a multi-center prospective ED cohort study to assess the inter-rater reliability of the mortality risk score derived in Specific Aim 1.